Machine learning methods and systems for identifying patterns in data using a plurality of learning machines wherein the learning machine that optimizes a performance function is selected

Abstract

Methods for training machines to categorize data, and / or recognize patterns in data, and machines and systems so trained. More specifically, variations of the invention relates to methods for training machines that include providing one or more training data samples encompassing one or more data classes, identifying patterns in the one or more training data samples, providing one or more data samples representing one or more unknown classes of data, identifying patterns in the one or more of the data samples of unknown class(es), and predicting one or more classes to which the data samples of unknown class(es) belong by comparing patterns identified in said one or more data samples of unknown class with patterns identified in said one or more training data samples. Also provided are tools, systems, and devices, such as support vector machines (SVMs) and other methods and features, software implementing the methods and features, and computers or other processing devices incorporating and / or running the software, where the methods and features, software, and processors utilize specialized methods to analyze data.

Description

RELATED APPLICATION DATA[0001]This application claims priority under 35 U.S.C. §119 to U.S. Provisional Application No. 61 / 095,731 filed on Sep. 10, 2008, the contents of which are incorporated herein by reference in their entirety.BACKGROUND OF THE INVENTION[0002]1. Field of the Invention[0003]Aspects of present invention generally relate to methods and systems for training machines to categorize data, and / or recognize patterns in data, and to machines and systems relating thereto. More specifically, exemplary aspects of, the invention relate to methods and systems for training machines that include providing one or more training data samples encompassing one or more data classes, identifying patterns in the one or more training data samples, providing one or more data samples representing one or more unknown classes of data, identifying patterns in the one or more of the data samples of unknown class(es), and predicting one or more classes to which the data samples of unknown clas...

AI Technical Summary

Benefits of technology

This patent describes methods using machine learning to identify patterns in data. The methods involve training multiple learning machines to identify which patterns are in the data samples. The learning machines are selected based on a performance function that maximizes the difference between different data groups, validates the training with cross-validation, and selects the most important variables. The trained learning machines can then be used to query data databases to identify patterns of data. The technical effect of this patent is to provide efficient and accurate methods for identifying patterns in data using machine learning techniques.

Problems solved by technology

Researchers are often unable to adequately query this data because they lack proper tools.

Currently, only a limited number of methods are available for searching across diverse gene expression experiments.

These experiments may include, for example, a study of diseased vs. normal tissue, the exposure of tissue culture cells to chemical compounds, or relationships between expression patterns in diseased cells and normal expression in other types of cells and search methods may be difficult to implement, even where these experiments were conducted using the same type of array format.

While SVMs and other machine learning tools have gained popularity in recent years, the methods remain primarily used for microarray analysis and classification, and have not been fully developed to optimize trained machines for searching and querying.

Current searching methods still rely largely on annotations, descriptive information, or values ranges for specific fields, and large amounts of data are therefore often not being utilized.

In particular, the choice of which trained machine is most likely to generalize well to data of unknown category is often a difficult one.

Also, a determination of which features in the data are the most important, and how many and which should be used in training is normally a complex problem that is challenging even to those well skilled in the art.

These systems and methods provide complex data processing, but rely upon inefficient techniques, such as expansion of training data and selection of optimum machines based on test data sets.

The ability of a learning machine to discover knowledge from data is limited by the type of algorithm selected.

Methods are also lacking that allow for the creation of hypothetical patterns and searching data to see if such patterns, or similar patterns, exist in actuality.

Images